[1] 11w IMPAIRED SYNAPTIC POTENTIATION PROCESSES IN THE HIPPO-CAMPUS OF AGED, MEMORY-DEFICIENT RATS

[1] 217w A series of neurophysiological experiments was performed on the Schaffercommissural system of the hippocampus of aged and young anesthetized Fischer rats. The aged Fischer rats were previously found to exhibit retention performance deficits. No obvious differences were found between aged and young animals in amplitude, latency, stimulation threshold, or wave forms of typical synaptic responses when these were elicited by control (0.3 Hz) stimulation pulses. Further, the temporal curves of facilitation during a paired-pulse series were not different in aged and young animals. However, aged and young synapses showed consistently different responses during repetitive stimulation. Synapses of aged animals were deficient in frequency potentiation processes during 12 Hz stimulation; and the aged animals exhibited a delayed rise of post-tetanic synaptic potentiation following a 5 sec, 100 Hz stimulation train. Moreover, aged synapses 'exhausted' more rapidly during continuous 4 Hz stimulation. Throughout these studies a biphasic pattern of potentiation was observed during repetitive stimulation (brief potentiation, depression, renewed potentiation). Aged animals were deficient primarily in development of the second phase of potentiation. This pattern suggests an age-related impairment of some secondary process of potentiation, leading to an increased tendency to synaptic depression during and after stimulation. The possibility that the impaired hippocampal synaptic plasticity may be related to reported deficient behavioral plasticity in the aged animals is considered.

[1] 164w Many histological and chemical correlates of brain aging have been reported in the literature10,1~,13,z0,45,47,49-51,55-57,a0-6L In contrast to the large number of reported neuroanatomical and neurochemical correlates of brain aging, however, there are almost no data available on the fundamental neurophysiology of the aging brain. Although a few basic neurophysiological studies of aging have been conducted, nearly all of these have employed spinal or peripheral systems 7,4s,63,64. A number of studies have examined age effects on the EEG and evoked potentials of animals and humans. These have indicated, in general, a reduced amplitude and frequency of synchronous waves and a reduced amplitude of evoked responses with age 5,1s,:19,44. Moreover, there is some evidence, as well as considerable speculation, that certain aspects ol synchronous EEG waves, including amplitude, may be correlated with complex memory and information processes (see review in ref. 30). Nevertheless, EEG studies of aging, while extremely valuable for correlative and diagnostic purposes, do not provide much information on possible cellular mechanisms of dysfunction.

[2] 97w At the other end of the age continuum, fundamental studies of hippocampal neurophysiology have provided important insights into the functional significance of the various anatomical patterns seen during development 46. To date, however, neurophysiological methods have not been used to aid our understanding of the functional consequences of anatomical and chemical patterns seen in the brain during aging. This lack of basic neurophysiological information regarding the aged brain seems to be a major obstacle to determining which, if any, of the brain's chemical and anatomical agechanges are directly relevant tofimctional decline, as has been noted elsewhere ~.

[3] 64w In an attempt to obtain this information, we have recently conducted a series of concomitant in vivo and in vitro studies on the hippocampal neurophysiology of aged and young Fischer rats. The in vivo studies are reported here, and studies using in vitro hippocampal preparations are reported in a separate paper 3"). Preliminary data from these studies have been described in abstract form 33.

[4] 130w The aged Fischer rats used in the present study were recently found by Gold and McGaugh zz to exhibit memory retention performance deficits. Another group has also found similar retention impairment in aged Fischer rats 45. These deficits seem to be analogous in temporal pattern to those in aged humans H,53 and in aged monkeys ' ~1 (i.e. the aged animals show normal acquisition, but retention performance decays rapidly over time). Moreover, the animals exhibit a distribution of brain deterioration 34 similar to changes seen for human cortex lz and for human hippocampus 4°,57. Thus, these Fischer animals appear to be reasonable subjects, not only for the study of the general neurophysiology of brain aging, but perhaps for the study of age-related mechanisms of changes in information storage as well.

[5] 195w The hippocampus was used in these investigations for a number of reasons. First, the distinctly laminated anatomical structure of the hippocampus makes it an excellent model cortical system4,8,23,2~.36,~7; second, the basic neurophysiology of the hippocampus is comparatively well understood as a consequence of extensive and thorough intracellular and extracellular analyses of the various major pathways 1-3,14,21,'~7,3~, 46,58; third, in humans 40 and in animals 12,34,47, the hippocampus is one of the structures affected earliest and most consistently by the histopathology of aging; therefore, any general brainwide physiological deficits that develop with age might well be detectable in hippocampus before they can be seen in other structures; fourth, in humans at least, the hippocampal region clearly seems to be linked to 'recent' memory processes 4'~, and human recent memory is consistently affected by aging well before the development of 'senility'Sa; fifth, hippocampal synapses exhibit one of the longest duration forms of post-tetanic potentiation 9,15, and one of the most robust forms of frequency potentiation 3 so far observed for any vertebrate synapse. This latter point seems relevant since investigators have long speculated about a possible role of synaptic potentiation processes in behavioral plasticity (e.g. ref. 17).

[6] 122w In the studies reported below, a number of experiments were conducted sequentially on each animal. This procedure may affect the absolute values of data obtained in the later experiments because of the influence of the earlier stimulation. However, this procedure also allows for the collection of a great deal of relative data, on which aged and young can be compared under a wide variety of conditions. Since the aged animals are extremely expensive, and difficult to raise in germ-free conditions, it therefore seems justified, so long as the treatment sequence is the same for all groups at comparable points in the same sequence. Instances in which the earlier experiments appear to have affected the later values have been noted in the text.

[1] 100w No significant differences were observed between young and aged animals in the following measures: threshold to elicit a population spike, amplitudes of population spikes at threshold or at various supra-threshold voltages, latency of response onset, stimulation voltage needed to elicit a 600/~V population EPSP, and wave forms of the responses. Additionally, there did not appear to be major differences in spontaneous single unit activity or amplitudes of unit spikes, although these were not systematically analyzed. Of course, there may well be small differences in threshold, input-output functions, spontaneous activity, etc., which might be found by more extensive and precise analyses.

[2] 50w No differences were observed in the time curves of paired-pulse facilitation, nor in the temporal course of decay or in the maximum points of facilitation (Fig. 2). However, a non-significant tendency for the amplitudes of the second EPSPs to be larger at briefer paired-pulse intervals was observed in young animals.

[3] 203w Significant differences were found, however, in nearly all experiments employing repetitive stimulation. During frequency potentiation (FP) both aged and young animals exhibited initial spike potentiation, often associated with the development of double or triple population spikes. In young animals, frequency potentiation tended to increase with increasing stimulation duration (at least within these parameters); while in older animals, an increased duration of stimulation was associated with decreasing population spike amplitudes. With higher frequencies (e.g. 12 Hz vs 6 Hz), the population spike decrease in aged animals occurred more rapidly. Thus, as can be seen in Fig. 3, the age-related impairment of frequency potentiation mechanisms appears to be both frequency-and duration-dependent. A biphasic pattern of response was often observed during the frequency potentiation series, particularly among young animals: the spike showed initial potentiation, then some depression, usually followed by a renewed and vigorous potentiation. This 'second phase' of potentiation appears to be the basis for the higher values of FP found in young animals at 15 sec as opposed to 5 sec. Moreover, this pattern seems to suggest the presence of some sort of delayed 'mobilization' process at hippocampal synapses; this process is seemingly activated approximately 5-15 sec after the onset of 6-12 Hz stimulation.

[4] 435w As seen in Fig. 4, 100 Hz stimulation for 5 sec elicited a large post-tetanic potentiation (PTP) in young animals. This PTP was also somewhat biphasic, but with a d!fferent time course. Immediately after stimulation termination, a large potentiation was observed, which rapidly decreased to a depressed or only slightly potentiated level by 10 sec poststimulation. This initial large PTP was extremely transitory and difficult to measure accurately, and is not shown in Fig. 4. During the early depressed phase that A large early potentiation in both groups was observed immediately following 100 Hz stimulation, but this rapidly decayed to a depressed stage between 10 sec and 1 min. Because of the rapid decay, this large early potentiation apparently lasting less than 10 sec is not seen in the graph. The only significant differences (asterisks) between aged and young were found in the Schaffer collateral system at 2 and 5 min post-100 Hz stimulation, and appear to reflect a more rapid recovery in the young animals from the depression that followed the early potentiation. Mean _~ S.E.M. followed initial PTP, the Schaffer collateral population spike sometimes disappeared. The population spike, however, generally recovered most of its potentiated level by 2 min poststimulation in younger animals. Spike recovery or growth tended to continue between 5 and 15 min, at which time it usually became asymptotic. Other investigators have noted that potentiated hippocampal region responses generally remain elevated at asymptotic levels for hours or days 9,15,~2, and we have observed similar results (numerous unpublished observations). Aged animals also exhibited an initial large PTP, followed by depression at 10 sec. However, in aged animals the recovery of PTP appears to be considerably delayed. Significant differences in PTP in the Schaffer collateral system between aged and young animals were found only during the early phases (Fig. 4). The PTP of the aged animals generally rose to the level of PTP in young animals by 15-30 min, although the largest degrees of PTP seen in some young animals (e.g. 60(0700 ~) were not attained by any aged animals (e.g. see means and S.E.M.s in Fig. 4). Thus, studies with larger groups might well indicate that the eventual levels of PTP attained by aged animals are also different from young animals. Fig. 5 shows that continuous 4 Hz stimulation also induced significant differences in the synaptic responses of aged and young animals, beginning approximately 45-60 sec after stimulation onset. Continuous 4 Hz stimulation depressed the Schaffer collateral population spike of most aged animals within 5 min. In younger animals, the response was generally stable for 5 rain of 4 Hz stimulation.

[5] 109w In this study also, a biphasic pattern was often seen. Following an initial potentiation to 4 Hz, an initial small decreasing phase was often observed, followed by a recovery and then stabilization. A similar depression and an attempted recovery were also seen in a number of aged animals, but this recovery was usually abortive, leading to further depression, or to stabilization at lower levels (Fig. 5). Thus, there appears to be a greater tendency to 'exhaustion' or 'fatigue' at aged synapses during continuous stimulation. (No initial frequency potentiation is seen in Fig. 5 because the control spike amplitude was measured from an already potentiated level, following the PTP series.)

[6] 98w In the recent literature, it has been reported that monosynaptic 'habituation' occurs at a hippocampal synapse 5s and neural habituation has been extensively studied as a process conceptually different from exhaustion 5a. Therefore, attempts to differentiate these processes were made. At the conclusion of the 5 min measurements ] Exhaustion / _~ young-mature 120-I [ ~; aged \ • 0 0 80-70-~ 60 n e-40 o 30-:~ 2o-I0-0 o ,~ 3'0 4'5 e'0 9'0 ,~o ,~o a~o 360 Seconds of Stimulation Fig. 5. Effects of continuous 4 Hz stimulation on the amplitude of the Schaffer collateral population spike.

[7] 204w After approximately 45-60 sec of stimulation, the amplitude of responses at aged synapses generally decreased rapidly. After a depression phase, a small recovery developed in aged animals, which, however, typically failed. In a number of young animals exhibiting a similar brief depression phase, a recovery also occurred. However, in young animals, the recovery was generally sustained thereafter for the duration of stimulation. The Schaffer collateral population spike was depressed essentially to zero by 5 min in many aged animals (asterisks reflect significant differences). Frequency potentiation is not seen during the early phases of 4 Hz stimulation in young animals because control responses for this 'exhaustion' series were obtained following the post-tetanic potentiation series and, therefore, reflect an already potentiated response (see text for procedures). Means 4-S.E.M. for 'dishabituation' were run in aged animals. Such procedures are generally viewed to be acceptable criteria for differentiating exhaustion from neural habituation.~S..~L Varying the frequency of stimulation did not produce 'dishabituation' (recovery) nor did increasing or decreasing the stimulation intensity. At very high intensities, the spike could sometimes be elicited again, but often at significantly reduced amplitudes in comparison to control levels. Thus, this depression phenomenon appears to be due t~ fatigue processes, rather than to habituation-like phenomena.

[8] 48w The spike amplitudes exhibited substantially greater variance among young than among aged animals in most of the experiments, primarily because of extremely large potentiation in some young animals. Moreover, for similar reasons, variability often changed substantially during stimulation. These effects are shown in Figs, 3, 4 and 5.

[9] 111w in young animals, repetitive stimulation often elicited double or triple population spikes to each stimulation pulse during frequency potentiation, as noted by other in-vestigators1,3, ~'~, with spikes separated by 4-8 msec intervals. This occurred typically during large frequency potentiation, and was also seen briefly in many aged animals, prior to the onset of spike depression. Potentiation of CA 1 responses has been described in several papers, and the patterns seen here were similar to prior reports, in cases of depression, the spike disappeared or was reduced, the extracellular 'IPSP' was proportionately reduced, and the 'EPSP' appeared basically unchanged in wave shape. Eatency and rise-time were not systematically studied following repetitive stimulation.

[10] 114w Control experiments with antidromic activation of CAI by stimulation of the alveusl, ' ~5 induced no age-related differences in CA1 responses during many minutes of alvear stimulation at frequencies up to 30 Hz, nor was consistent potentiation or depression seen during many minutes of such stimulation. At higher frequencies, (e.g. 30 Hz) the antidromic response decreased up to a third in amplitude, possibly because of increased recurrent inhibition 2,27. However, responses were stable at this level and no differences were seen between aged and young animals: in no case was complete spike depression observed, which did, however, occur often in aged animals at frequencies above 10 Hz using orthodromic stimulation of the Schaffer collaterals.

[11] 72w It is possible that the fiber properties of the Schaffer system, which was activated orthodromically, are different from those of the alveus, which was used for antidromic control studies. However, both systems are myelinated, and in other studies, performed in vitro ~2, similar age-related orthodromic effects were obtained, but antidromic sti mulation of both alvear and Schaffer fibers produced no age differences. Thus, these effects seem to be limited to orthodromic activation.

[12] 122w Histological analysis showed that in all animals examined, the Schaffer collateral stimulation electrode was in the CA3 cell layers or in the dendritic region in which the Schaffer collaterals travel. The CA 1 recording pipettes were also found to be correctly oriented in all animals studied. The pipette tips could not be localized with these methods, but our physiological criteria for localization in the cell layers and dendrites closely agreed with findings from other studies in which extensive physiological-anatomical correlations have been performed in the CAI field 1,~1, and also agreed very well with stereotaxic coordinates ~8. No age-related differences in the histological localization of the electrodes were observed, and the variation appeared to cover the same anterior-posterior coordinates in both groups.

[1] 159w The results reported here indicate that the hippocampal physiology of aged animals differs from that of young animals primarily when the synapses are 'challenged' by repetitive stimulation. Synaptic responses of aged animals to control (0.3 Hz) stimulation appeared similar to those of young-mature animals with regard to latency, wave-form, thresholds or response amplitudes at 20 ~ above spike thresholds. More detailed studies of basic physiological responses to single-pulse stimulation are planned for the future to determine whether small age differences in threshold, conduction velocity, or input-output functions are present. However, it seems clear that aged and young hippocampal synapses are considerably more different during repetitive stimulation than during what is essentially single-pulse (i.e. 0.3 Hz) control stimulation. Thus, it seems to be the neural responses which are generally elicited by repetitive stimulation (e.g. potentiation, depression) that are most changed in the aged hippocampus; this seems of some interest since these responses have often been speculatively linked to behavioral plasticity.

[2] 155w The observed age-related effects of repetitive stimulation are, as noted, based upon sequential studies. Thus, even though the two age groups were subjected to the same sequences, it is of course possible that the earlier stimulation affected the aged animals more than it did the young. These results, therefore, indicate an age-related effect of the overall sequence of repetitive stimulation, but cannot be conclusively linked to any single experimental paradigm. Nevertheless, the groups were comparable at the slart of each experiment, and the aged animals exhibited normal early frequency potentiation and normal responses during the early phases of the exhaustion experiment. This suggests that prior experiments may not have greatly influenced the results of the later ones; moreover, it should perhaps be noted that we have obtained similar results, with regard to frequency and post-tetanic (or long-term) potentiation in in vitro studies of the hippocampal slice in which only one stimulus train was administered 32.

[3] 91w The CA1 field of young animals in the present study appeared to show synaptic response patterns similar to those reported for rabbits 1, cats 21, and other strains or species of rodentsa4,39, ~. However, the percentage increase in spike amplitude during potentiation was somewhat less in the present study than is generally reported 1,35,3.9,52. This can probably be accounted for by the substantially larger control level of spike amplitude used in this study (e.g., Fig. 1 vs Fig. 6 in ref. 52). This factor can greatly influence percentage spike changesaS, 5~.

[4] 118w As described in the Results section, potentiation mechanisms at hippocampal synapses appear to involve a biphasic process, both during and after stimulation. The time course of the development of the second phase seems to depend upon stimulation frequency, and seemingly suggests the presence of a delayed, secondary process of potentiation which may be different from that elicited by the first pulses. It seems conceivable that the particularly robust potentiation seen in hippocampus is at least partially due to this secondary, delayed potentiation process. The aged animals, moreover, primarily appear deficient in the development of this second phase. An understanding of the mechanistic significance of this second phase of hippocampal potentiation, however, seems likely to require considerable further research.

[5] 95w The apparent failure of some aspects of synaptic potentiation in aged animals. however, may not be due solely to a deterioration of potentiation mechanisms per se. That is, the initial levels of frequency potentiation during repetitive stimulation (Fig. 3) and the eventual levels of posttetanic potentiation attained (Fig. 4) by aged animals~ were not significantly different from levels seen in young animals. These findings could therefore suggest that the mechanisms of potentiation of aged synapses are not greatly impaired, but that instead, the aged synapses may be more sensitive to depressive effects of repetitive stimulation.

[6] 97w At many synapses, stimulation induces two independent processes of potentiation and depression e6,29. Both processes also appear to occur at hippocampal synapses (see discussion in ref. 35). Nevertheless, it seems far from certain that these processes are essentially the same in hippocampus as in other systems. The extremely long-lasting, nearly semi-permanent potentiation seen in hippocampus is different from the potent iation in most systems studied, which generally decays in minutes or hours. Moreover, recent findings suggest that low-frequency hippocampal depression may involve some postsynaptic aspects 3s,39, whereas depression seems to be presynaptic in most other systems studied.

[7] 56w However, the impaired potentiation or depression seen in the present study in aged animals is more pronounced at higher frequencies (Fig. 3), and therefore appears different from the low-frequency depression seen in young animals 3s,39. In some ways, then, the age-related deficits observed here appear more analogous to studies in other systems involving presynaptic processes 26,'~9.

[8] 88w Although substantially more research will clearly be required to establish the preor postsynaptic basis of the age-related synaptic deficit reported here, there does seem to be a reasonable possibility that the deficit may involve presynaptic transmitter processes, since depletion of synaptic vesicles (and depression) is also increased by increased frequency and/or duration of stimulation in several vertebrate synaptic systems z4,43. Quantal analyses of various defined synaptic systems have suggested that facilitation and potentiation may involve transmitter mobilization whereas depression may involve depletion of the releasable transmitter pooF 6,'~9,6'~.

[9] 142w Other data strongly suggest the possibility that transmitter functions may be deficient in aged animals. A decreased metabolism of catecholamines as well as reduced catecholamine levels in some regions, has been reported in the brains of aged animals 20. 55. Alterations in cholinergic and in putative amino acid transmitter metabolism have also been noted6°,6L Moreover, a decreased frequency of spontaneous miniature end plate potentials has been reported at neuromuscular junctions of aged animats 63. A possible impairment of axonal transport in aged brains has also been proposed :'7, which could lead to decreased vesicle populations. There is additionally, a clear degeneration of synaptic elements in aged brains 1°,51,57,61. Thus, at least to this point, the present physiological data seem consistent with neurochemical and morphological evidence indicating some form of impairment of transmitter or other synaptic functions in the brains of aged animals.

[10] 104w Impaired synaptic potentiation in the hippocampus of aged rats, regardless of the cellular mechanisms involved, could conceivably influence the efficiency of in-formation processing or storage. We have suggested elsewhere 32,a3 that impaired hippocampal synaptic potentiation may be responsible for the 'short-term' memory deficits observed in these aging animals. In humans the relationship of the hippocampal regions to 'short-term' memory processes appears supported by considerable data. Although the functions of the hippocampus in non-human mammals are controversial, recent findings have shown that experimental driving of the characteristic hippocampal theta rhythm, utilizing septal stimulation at 7.7 Hz, can improve apparent post-trial memory processes in rats 30,31.

[11] 57w Nevertheless, the view that we have proposed --that the impaired potentiation is a significant factor in the age-dependent memory decline in Fischer rats --is a highly tentative hypothesis which we suggest primarily to provide a clear basis on which to test this possibility. Hopefully, this may also facilitate the general analysis of age-related decline of brain function.

[1] 177w The experiments were performed using inbred, male, albino Fischer rats, barrierreared by Charles River Co. under an agreement with the National Institute on Aging. Some of the animals had first been used briefly in a series of behavioral experiments in which the aged Fischer rats have been found to show retention deficits ~2. Until the animals were used for physiological experiments, they were maintained in an isolated clean room apart from the other animals, and their cages were changed only by handlers in clean coats and gloves. Young-mature animals were 4-7 months old, and aged animals were 25-27 months old when used. Data were obtained from 11 aged and 13 young animals. Of these, 9 young and 9 aged animals were carried throughout the entire sequence of 4 experimental series (see below), and approximately half of each group was then used for antidromic studies. Additionally, 2 aged and 4 young animals were employed solely for antidromic control or localization studies. Aged animals showing any overt signs of pathology (e.g. tumors, weight loss, infection) were excluded from study.

[2] 66w Animals were anesthetized with intraperitoneal injections of urethane (l g/5 ml/kg) and were occasionally given supplemental doses of 0.1-0.2 g/kg when needed. Local anesthetics (procaine) were standardly applied to wound edges. The animal was placed in a heavy stereotaxic instrument with the skull flat; the skull was surgically exposed, and holes were drilled in the skull with a small dental drill at appropriate coordinates (see below).

[3] 81w The same basic series of experiments was performed on the Schaffer collateral system of each animal. This system also includes commissural fibers which travel with the Schaffer collaterals in stratum radiatum 2a (for purposes of convenience, however, these fibers will generally be referred to here as the Schaffer collateral system). Standard electrophysiological recording and stimulating equipment and procedures were employed, and responses were displayed on a Tektronix storage oscilloscope for amplitude measurements. The indifferent electrode was grounded to the stereotaxic frame.

[4] 203w Electrode placements were made based upon both physiological criteria and stereotaxic coordinates. All electrodes were inserted using micromanipulators (Kopf) with minimal calibrated increments of 10 #m. The recording micropipette (tip diameter: 1-3 #m; impedance: 3-8 M~; 2.5 M NaC1) was vertically lowered into CA1 of the left hippocampus at the following coordinates: 3.5-4.0 mm posterior to bregma: 2.0 mm lateral to the midline. At approximately 2.0-2.2 mm below the brain surface, large single unit spikes were recorded from multiple cells in the pyramidal cell soma layer CAl. These spikes generally showed typical pyramidal cell patterns of discharge (i.e. large amplitude, double spike discharges, with spikes separated by several msec) TM 46. Single unit spikes are primarily only recorded from the cell layers within the hippocampus with these methods. The pipette tip remained in the CAI cell layer while the stimulating electrode (bipolar; 2 twisted strands of 120/~m enamel-coated stainless steel wire, evenly cut at the tips; 20-30 kf~ DC resistance) was lowered into the Schaffer collaterals at the following coordinates: 3.5-4.0 mm posterior to bregma, 4.0 mm lateral to the midline, and angled 15 ° towards the midline. Beginning 2 mm below the surface, the typical Schaffer collateral synaptic response was generally observed.

[5] 211w This response has been extensively analyzed in detail by Fujita and Sakata zl and by Andersen and his associates 1, and only a brief description will be given here. In the cell layer a slow positive potential is seen with an onset latency of 2-4 msec. This is similar to the response observed in other hippocampai fields 35. If the stimulation voltage is sufficiently intense, a large negative population spike is seen riding on the slow positive wave, and this is immediately followed by a larger slow positive wave (Fig. 1). The initial slow wave seems to be the extracellular summated population EPSP; the sharp negative population spike has been shown to be an 'envelope' of summated single unit spikes, and the subsequent large positive wave appears to be a summated extracellular IPSP, which is triggered by the cell firing (and recurrent inhibition) associated with the population spike 1,2,27. The EPSP reverses polarity and becomes negative as the micropipette is lowered approximately 200 #m beyond the layer of the pyramidal somata into the apical dendrites1,21, 35, as is shown in Fig. 1. The most likely explanation of this reversal is that the cell somata act as a 'source' for the active 'sink' on the dendrites where the Schaffer collaterals terminate1, 2~,32.

[6] 48w Once a CA1 response was obtained which met our criteria for Schaffer collateral activation (onset latency under 3.2 msec, polarity reversal of the EPSP when the pipette was lowered), the recording electrode was returned to the cell layer and was maintained there for the duration of the experiments.

[7] 161w The threshold for eliciting the population spike was initially determined during 0.3 Hz stimulation (general range: 300-800 #A). The voltage and current (across a 1 kf~ series resistance) were simultaneously monitored in most instances. Current was generally stable, but occasional small fluctuations were compensated for by very minor voltage adjustments. The voltage was then lowered to approximately 50 ~ of the threshold value in order to conduct a paired-pulse facilitation series. The lower voltage was necessary because when a population spike is elicited, the spike response to the second stimulation pulse is often inhibited for 30-100 msec, a finding observed even when microelectrode stimulation of the Schaffer collaterals is employed2L This inhibition appeared CAI Rec. fffer m. to be less pronounced in in vitro hippocampal slice preparations3L Below spike threshold, only the slow population EPSP is elicited (positive in cell layer, negative in apical dendrites), and the inhibition period is much reduced. Spike-induced inhibition is presumably due to recurrent inhibition 1,2,~7.

[8] 83w Using this voltage level, responses were measured during 0.3 Hz stimulation. At this stimulation frequency responses are highly stable, and the response amplitude at 0.3 Hz was taken as the control value. The stimulator was then set in the paired-pulse mode and the interval between the first and second pulses was systematically varied between 20 and 200 msec. The amplitude of the second response was measured at 20. 30, 40, 60, 80, 100, 120, 140, 160, 180 and 200 msec intervals (positive peak-to-baseline).

[9] 44w At each interval, two sets of paired pulses were given at 0.3 Hz, and the average of the two responses to the two second pulses was taken as the measured value at that interval. Very little variation of responses occurred at a given duration.

[10] 272w Following the paired-pulse series, the voltage was raised to approximately 20'J,, above the level needed to elicit a fully developed population spike. In some cases, inputoutput relations were briefly studied. Control values were then obtained at 0.3 Hz stimulation. (Measurements of spike amplitude were always made from the origin of the spike on the EPSP to the negative peak of the spike.) Frequency and post-tetanic potentiation series were all conducted at this stimulation intensity. Frequency potentiation (growth of a response during repetitive stimulation) was measured by stimulating at 6 Hz for 5 sec. After a I min interval, during which 0.3 Hz control stimulation was administered, a second series was performed, stimulating at 6 Hz for 15 sec. After a 2 rain interval of control stimulation, a third series was also conducted, stimulating at 12 Hz for 5 sec, and after a 1 min interval of control stimulation, a fourth series at 12 Hz for 15 sec was performed. With the exception of the last series, responses of both groups to 0.3 Hz following each series were similar to control levels. The last two synaptic responses at the end of each of the four series were displayed on the storage oscilloscope, and the average of the two responses (which varied very little) was taken as the frequency potentiation value (in comparison to initial control values) for that series. Although many components of the response exhibit potentiation, the greatest change is found in population spike amplitude 35,52. Therefore, in the Schaffer collateral system, amplitude of the population spike was used as the basic measure during the frequency potentiation, post-tetanic potentiation and 'exhaustion' studies.

[11] 211w A 5 min interval was allowed between the last frequency potentiation series and the start of the PTP series. During this interval, 0.3 Hz stimulation was given. For animals that did not show spike values at control levels during this interval, voltage was adjusted slightly to produce the initial control values. Five minutes after the last frequency potentiation series, post-tetanic potentiation (PTP --growth of response to control pulses following termination of repetitive stimulation) was studied by administering a 100 Hz stimulation train for 5 sec. (This terminology of'post-tetanic potentiation' is strictly descriptive; it implies nothing regarding the pre-or postsynaptic mechanistic basis of the phenomenon.) This intense (100 Hz) stimulus was employed in order to overwhelm any effects of the prior 12 Hz stimulation series. The amplitude of the population spike to 0.3 Hz pulses was then measured at 10 sec, 2 min, 5 rain, 15 rain and 30 min following the 100 Hz train. At each measurement interval, the average amplitude of three responses was taken as the value for that interval. The three responses typically varied very little except at the 10 sec interval, at which point the first response was generally extremely large, with the second and third responses decreasing very rapidly, often to the point of total depression.

[12] 151w Studies of'synaptic exhaustion' were initiated following the 30 rain measurements of PTP. A new control value for the population spike was obtained during 0.3 Hz stimulation, due to the prior influence of tetanizing stimulation. These values were obtained at the 30 min PTP period at which point the two groups were not significantly different (Fig. 4); therefore, in terms of the initial pre-frequency potentiation control values, the groups were still roughly comparable. Exhaustion studies were carried out by administering continuous stimulation at 4 Hz for 5 min. Averages of two synaptic responses were obtained at 15, 30, 45, 60, 90, 120, 240 and 300 sec after the onset of 4 Hz stimulation. No interruption of stimulation occurred during these measurements. An average of two measurements was sufficient since these responses are highly stable over short periods of time (e.g. 5 sec), even though they may be changing substantially over longer periods.

[13] 59w In several animals, lowering the stimulation electrode induced an electrographic seizure and subsequent depression of spontaneous slow wave amplitudes and single unit activity. In these cases, which occurred with approximately equal frequency in aged and young animals, an interval of 15-20 min was allowed before continuing. By that time, slow wave amplitudes and spontaneous unit spikes were fully recovered.

[14] 89w In 7 aged and 7 young animals, antidromic control experiments were performed. In these, the pipette was maintained in the CA1 pyramidal layer, but the stimulation electrode was aimed at regions of the alveus in which run CA1 output fibers1, zS. What appeared to be an antidromic response (i.e. 0.5 msec latency, negative spike not preceded by a slow wave, and followed by a positive slow wave) could often be elicited using these placements. Experiments with repetitive stimulation similar to those carried out orthodromically were performed with these responses.

[15] 44w At the completion of an experiment, approximately half of the animals were perfused intracardially with saline and 10 ~ formalin, and the brains were removed and stored in formalin. They were subsequently frozen, sectioned (35/zm) and stained (cresyl violet) for verification of electrode placements.

[16] 159w Data analysis was performed by converting all responses to percent of the initial control values (obtained prior to the facilitation, frequency potentiation, and exhaustion series). Data were analyzed using both parametric and non-parametric statistics. However, tests showed that in most cases the assumption of homogeneity of variance had to be rejected, and a normal distribution could also not be assumed, since many aged animals exhibited response depression to zero. Therefore, statistical inferences were based largely on distribution-free statistics. However, it should be noted that parametric analyses of variance and individual group comparisons showed highly similar results, Mann-Whitney U tests were used to compare aged and young animals at various time points, and Wilcoxon matched-pairs, signed-ranks tests for paired samples were used to determine whether a change within a group was significantly different from baseline 54. All reported effects are significant at least at the P < 0.05 level, two-tailed, and group differences are represented in the figures by asterisks.

[17] 99w In brief, then, both the aged and young groups were carried through precisely the same sequence of 4 experiments (facilitation, FP, PTP, exhaustion). At the start of each experiment, the two groups were not significantly different from each other during 0.3 Hz stimulation (in terms of percentage changes from original control values). In only one experiment--exhaustion --did each group begin significantly above its initial control levels (as a consequence of the prior PTP). Nevertheless, the two groups did not differ significantly from one another. A new control value to 0.3 Hz stimulation was taken prior to the exhaustion series.